XDOF, just three months out of stealth, is in talks for a Series B at a $1.2B valuation

Robotics data startup XDOF is in late-stage negotiations to secure a Series B funding round at a valuation of approximately $1.2 billion, with 8VC reportedly leading the investment. This development comes less than three months after the company emerged from stealth and announced its Series A, highlighting intense investor appetite for foundational infrastructure in the physical AI sector. The rapid succession of funding rounds for a company founded just this year indicates that the market sees a critical need for specialized data providers to overcome the primary bottleneck hindering the development of general-purpose robots: the scarcity of large-scale, real-world interaction data.

The potential new funding follows a $70 million Series A in June, which included participants like Thrive Capital, Andreessen Horowitz, and Spark Capital. While XDOF wasn't actively seeking new capital, its swift operational ramp-up, with annualized revenue reportedly nearing $50 million, attracted inbound interest from venture firms. The company, co-founded by UC Berkeley researchers Philipp Wu and Fred Shentu, is building a sophisticated data supply chain for the robotics industry, addressing a problem Wu encountered directly during his PhD research—the lack of sufficient data to train capable robots.

XDOF's Data Collection Infrastructure

XDOF's technical approach is rooted in the founders' research on a low-cost teleoperation system called GELLO. The company employs a hybrid model to generate the high-quality datasets its clients, including several frontier AI labs, require.

• Remote Teleoperation: Human operators remotely control robotic arms and other hardware to perform and record complex tasks, generating precise machine-centric data.
• Egocentric Data Capture: Human collectors wear body sensors to record movement and interaction data from a first-person perspective while performing everyday activities like folding laundry or assembling objects.
• Data Pipeline: The company provides the full stack of collection tools, data pipelines, and annotation systems, allowing robotics companies to outsource this complex but essential function.

This strategy positions XDOF as a direct enabler for the broader robotics market, drawing comparisons to how data-labeling firms like Scale AI fueled the growth of large language models. Unlike LLMs, which could be trained on the vast corpus of the public internet, physical robots require proprietary, structured data captured from the real world. By building a scalable solution for this, XDOF is creating a foundational asset for an entire industry. The company faces competition from other data-focused startups like Mecka AI and incumbents such as Scale AI and Micro1, which are expanding their services beyond LLM data into the physical domain.

Strategic Takeaway: XDOF's aggressive $1.2 billion valuation reflects a market consensus that proprietary, real-world physical interaction data is the most critical and defensible asset for building general-purpose robots. Investors are not just funding a data-labeling service; they are backing a company that aims to own the foundational data supply chain for the entire physical AI economy, making its platform a prerequisite for any serious player in the field.